Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
import type { Tokens, Lexer, TokenizerExtension } from 'marked'
export type ContainerOptions = Record<string, boolean | string>
export type TokenExtractionParameters = {
/**
* The type of container
*/
type: string
/**
* The content inside the container
*/
content: string
/**
* The raw full container block
*/
raw: string
/**
* The options of the container
*/
options: ContainerOptions
/**
* The lexer that can be used to render nested tokens if needed
*/
lexer: Lexer
}
export type TokenExtractor = (
params: TokenExtractionParameters,
) => Tokens.Generic | null
/**
* Container extension is a marked extension that parses a block of text surrounded by :::
*
* @param tokensExtractor Converts the container input into a custom token that svelte components will receive.
* If this function returns a value that contains an array of `tokens`,
* this array will be populated with the tokens that are found inside the container,
* allowing to nest markdown inside the container.
*
* @example
*
* Given the following markdown:
*
* :::container option="option-1" option-2="option-2" boolean-option
* Something inside the container
* :::
*
* The token extractor would receive the following parameters:
* **type**: `container`
* **content**: `Something inside the container`
* **options**: { 'option': 'option-1', 'option-2': 'option-2', 'boolean-option': true }
*/
export default function (tokensExtractor: TokenExtractor): TokenizerExtension {
return {
name: 'container',
level: 'block',
start(src: string) {
return src.match(/^:::[^:\n\s]/)?.index
},
tokenizer(src: string): Tokens.Generic | undefined {
const rule =
/^:::(?<type>[a-z0-9-]+)(?<options>.*)?\n(?<content>(?:.|\n)*)\n:::(?:\n|$)/i
const raw = findRawContainer(src)
if (!raw) return undefined
const match = rule.exec(raw)
if (!match || !match.groups || !match.groups.type) return undefined
const type = match.groups.type.toLocaleLowerCase()
const options = parseOptions(match.groups.options || '')
const content = match.groups.content || ''
const result = tokensExtractor({
type: type,
content: content,
raw: match[0],
options: options,
lexer: this.lexer,
})
if (result && result.tokens) {
this.lexer.blockTokens(content, result.tokens)
}
return result ?? undefined
},
}
}
function findRawContainer(src: string): string | undefined {
if (!src.startsWith(':::')) return undefined
const lines = src.split('\n')
let open = 1
let lineNumber = 1
for (lineNumber = 1; lineNumber < lines.length; lineNumber++) {
const line = lines[lineNumber]
if (line && line.startsWith(':::')) {
if (/:::[^:\n\s]/.test(line)) {
open++
} else if (/^:::(\n|$)/.test(line)) {
open -= 1
}
}
if (open === 0) {
break
}
}
return lines.slice(0, lineNumber + 1).join('\n')
}
function parseOptions(options: string): ContainerOptions {
const output: ContainerOptions = {}
let remaining = options
while (true) {
const regex = /(?<name>[a-z0-9]+)(?:="(?<value>(?:.*?))")?/i
const match = regex.exec(remaining)
if (!match) break
const name = match?.groups?.name
if (name) {
output[name] = match?.groups?.value ?? true
}
remaining = remaining.slice(match.index + match[0].length)
}
return output
}